From Algorithmic Black Boxes to Adaptive White Boxes: Declarative\n Decision-Theoretic Ethical Programs as Codes of Ethics

2017/11/16 by Martijn van Otterlo, van Otterlo, Martijn
Social Sciences · Computer Science · #Ethics and Social Impacts of AI #Explainable Artificial Intelligence (XAI)

paper · pdf · doi:10.48550/arxiv.1711.06035

Abstract

Ethics of algorithms is an emerging topic in various disciplines such as\nsocial science, law, and philosophy, but also artificial intelligence (AI). The\nvalue alignment problem expresses the challenge of (machine) learning values\nthat are, in some way, aligned with human requirements or values. In this paper\nI argue for looking at how humans have formalized and communicated values, in\nprofessional codes of ethics, and for exploring declarative decision-theoretic\nethical programs (DDTEP) to formalize codes of ethics. This renders machine\nethical reasoning and decision-making, as well as learning, more transparent\nand hopefully more accountable. The paper includes proof-of-concept examples of\nknown toy dilemmas and gatekeeping domains such as archives and libraries.\n

Related

• From computational ethics to morality: how decision-making algorithms can help us understand the emergence of moral principles, the existence of an optimal behaviour and our ability to discover it
• On the Computational Complexity of Ethics: Moral Tractability for Minds and Machines
• Beyond Fairness: Alternative Moral Dimensions for Assessing Algorithms and Designing Systems
• Gatekeeping Algorithms with Human Ethical Bias: The ethics of algorithms in archives, libraries and society
• Practical Challenges in Explicit Ethical Machine Reasoning